Spatial Intelligence
This direction studies how large language models, vision-language models, and urban world models can understand cities, geospatial data, street views, satellite images, maps, trajectories, and multimodal urban signals.
Research Agenda
My work connects spatial intelligence, agentic intelligence, and urban science through models, agents, benchmarks, and real-world systems.
This direction studies how large language models, vision-language models, and urban world models can understand cities, geospatial data, street views, satellite images, maps, trajectories, and multimodal urban signals.
This direction builds LLM-powered agents and embodied agents that coordinate data processing, model planning, optimization, verification, and simulation for real-world applications.
This direction focuses on urban analytics and principled modeling of human behavior and urban systems, including trajectory prediction, urban flow forecasting, spatiotemporal learning, and industrial-scale decision systems.